Readiness score: what it actually measures

It is five to six in the morning. You reach for the phone before your feet hit the floor, and there it is. 41 out of 100. Red ring. "Take it easy today."

So now you are sitting on the edge of the bed negotiating with a number. Should you cancel the session? Your body feels fine, honestly. But surely the number knows something you do not?

It does not. It knows less than you think, and it knows it in a way worth understanding, because that understanding is the difference between a tool and a boss.

What the score is made of

In April 2025, the journal Translational Exercise Biomedicine published a review with an unusually plain purpose. The researchers wanted to know what these scores are actually assembled from. They worked through everything the manufacturers had made public: technical white papers, user manuals, the explanations inside the apps themselves, and whatever research existed.

They found 14 composite health scores across 10 of the largest manufacturers.

Then they counted the ingredients. Four measurements turned up more often than anything else.

Read that table twice, because it gives something away. There is no secret sensor. There is no measurement of "recovery" as an organ or a state. The score is built from numbers you can already see individually on the same screen.

What separates the manufacturers is not what they measure. It is how they weight and combine those measurements. And that is where the review lands its uncomfortable sentence: none of the ten disclosed their exact algorithmic formula, and few provided empirical validation or peer-reviewed evidence supporting the accuracy or clinical relevance of their scores.

Why two devices disagree about one night

The most common frustration with these scores is not that they exist. It is that they do not agree. The ring says 78, the watch says 44, and you only slept one night.

Two things explain that, and they act at the same time.

The first is the raw material. In August 2025, Physiological Reports published a validation study in which thirteen healthy adults, six of them women, wore an ECG reference and several consumer devices at the same time while they slept. That produced 536 nights. Each device's nocturnal resting heart rate and heart rate variability were then compared against the reference.

The figures are mean absolute percentage error against ECG. Lower is better. Garmin was excluded from the resting heart rate analysis because of methodological inconsistencies in how the device reports the value, which is itself a reminder that two devices do not always measure the same thing under the same name.

Look at the right-hand column. The best sat at roughly 6 percent error, the worst at just over 16. The error at the bottom of the list is almost three times the error at the top, measured against the same ECG on the same nights. Which device you happen to wear is therefore part of the answer to what your number turns out to be.

The second explanation sits on top of the first. Even if two devices had measured identically, they would still produce different scores, because they weight the ingredients differently and draw them from different time windows. The blending is proprietary at all ten companies. Two numbers built from different raw material and different recipes cannot possibly land in the same place.

If you want to understand why the wrist is a harder place to measure than the chest, that whole question is covered in the guide to chest strap versus wrist heart rate.

The score looks backwards

This is where the real misconception sits, and it is built into the word. "Readiness" sounds like a forecast. A verdict on what you can handle today.

But go back to the ingredients. Heart rate variability from the night that just ended. Resting heart rate from the night that just ended. Sleep duration from the night that just ended. Activity from the day before. Every single ingredient is history.

So the score is a summary, not a prediction. It says roughly this: this night looked different from your recent weeks, in this direction, by about this much. That is genuinely useful. It is just not the same thing as knowing how your session will go.

The gap is most obvious when the low number has an obvious cause. You ate late. You had a glass of wine. You ran intervals at nine in the evening. The score is then telling you exactly what you already knew, with two decimal places of confidence. What a late session does to your night is covered in the guide to exercise and sleep.

There is one exception worth taking seriously. When the number is low with no cause you can point to, and it stays there for several mornings in a row, it has told you something you did not know. That is the only pattern in this kind of data that earns the right to change your week.

The cheapest instrument has the best evidence

Now for the finding that turns the whole question around, and that no manufacturer has any interest in advertising.

In 2016 the British Journal of Sports Medicine published a systematic review of how athlete well-being is best tracked over time. The researchers looked for studies that measured objective markers, meaning heart rate, blood markers and performance tests, alongside subjective ones, meaning what the person reported about mood, fatigue and perceived stress. Fifty-six original studies met the criteria.

Two results came out of it.

The first was that the objective and the subjective measures generally did not correlate with each other. They simply were not measuring the same thing.

The second was that the subjective measures reflected both acute and chronic training load with superior sensitivity and consistency compared with the objective ones. Subjective well-being typically got worse with an acute increase in training load and with chronic training, and improved when the load was reduced.

In other words: the question "how does it feel today?" tracked training load better than the instruments did.

Build your own readiness in three lines

The conclusion from the two sections above is not to bin the watch. It is to put the number where it belongs: as one witness among several, not as the judge.

Ask three questions every morning, before you look at the phone. Rate each answer from 1 to 5, where 5 is best.

1. How did I sleep? Your own experience, not the app's sleep score.
2. How is my energy right now? Before the coffee, not after.
3. How heavy does my body feel? Your legs on the stairs say more than you would think.

The total lands between 3 and 15. After a couple of weeks you have a normal range, and only then does the number mean anything. Exactly as with individual metrics, the information is carried by the deviation from your own baseline, never by the absolute figure. How to establish a baseline in practice is covered in the guide to HRV.

Then let the watch score be the fourth voice. Here is how to read the outcome.

Notice what is missing from the right-hand column. The word "rest" does not appear once on the single-day rows. A low morning is almost never a reason to skip a session. It is a reason to make the session easier, which is a different thing and a far smaller price. How often the body actually needs a full day off is covered in the guide to rest days between workouts.

Row two is the most important row in the table. When the number says 88 and you can barely tie your shoes, trust yourself. The 2016 review points in exactly that direction.

After 40: what actually changes

Two things make this kind of data trickier in midlife, and neither of them means your body is broken.

The first is that your week contains more disturbances that are not training. A child who wakes up. A deadline. An evening that ran late for reasons you did not choose. All of it lands in the same score as yesterday's run, and the score cannot tell them apart. A low number after a night with a sick child is not a training verdict.

The second follows from the first. The more disturbances your week contains, the more noise there is in the series, so you need more days before a pattern can be told apart from chance. That argues for patience with the numbers, and against reading every morning as a fresh verdict.

The practical consequence is simple. Look at the week, not the day. Four of seven mornings below your normal range means something. A single Tuesday rarely does. Your watch's VO2 max figure needs even longer than that: months rather than weeks, as the guide to what VO2 max is explains.

When it is not about training

There is a point where this page stops being a training page, and it is worth saying out loud.

Tiredness that does not lift is not a question for your watch. No readiness score can make that assessment, and none of them are designed to.

One more thing, since it sits close by: persistent daytime sleepiness despite enough time in bed has a common cause that no score will pick up. That one is covered in the guide to snoring and sleep apnoea.

What this means on a Monday morning

The summary fits in four lines.

The score is not a measurement but a blend of four numbers you already have, and the recipe is proprietary at every manufacturer. It looks backwards, at the night that has ended, not forwards at the day you still have. The margin of error in the raw material is enough to make two devices disagree about the same night. And the instrument with the best research behind it is free: the same three questions, every morning, written down.

Use the number as one witness among several. Never let it be the one that decides. A missed session is data, not a failure, and a red ring at five to six in the morning is not a verdict on you either.

Frequently asked questions

What does a readiness score actually mean?

It is a composite number your watch or ring calculates from a handful of measurements taken during the night that has just ended. A review published in Translational Exercise Biomedicine in 2025 catalogued 14 such scores across 10 manufacturers. The most common ingredients were heart rate variability in 86 percent of the scores, resting heart rate in 79 percent, physical activity in 71 percent and sleep duration in 71 percent. So the score is a summary of what has already happened, not a measurement of some new organ in your body. In practice it says: this night looked different from your own recent weeks, in this direction, by roughly this much.

Why do my watch and my ring show different readiness on the same morning?

For two reasons that act at the same time. The raw ingredients differ, and the recipes differ. In a validation study published in Physiological Reports in 2025, thirteen adults wore an ECG reference alongside several consumer devices while they slept, across 536 nights. For heart rate variability, the mean absolute percentage error ranged from 5.96 percent for the most accurate device to 16.32 percent for the least. On top of that, each manufacturer applies its own weighting, and none of them publish the formula. Two numbers built from different raw material and different weights cannot possibly land in the same place.

Should I skip training when my readiness is low?

Rarely entirely, often a little. A single low morning after a late session, a glass of wine or a broken night is expected, and it is not a stop sign. What is worth acting on is a number that stays low for several days with no cause you can point to, or a low number that matches how you actually feel. In practice that usually means keeping the session but lowering the ambition: swap the intervals for an easy run, drop a set, shorten the session. Removing training altogether is rarely the right answer to a number on its own.

Are readiness scores scientifically validated?

Not in the way the word suggests. The 2025 review found that none of the ten manufacturers disclosed their exact algorithmic formula, and that few provided empirical validation or peer-reviewed evidence supporting the accuracy or clinical relevance of their scores. That is a verdict on the evidence base, not proof that the number is wrong. The individual ingredients, especially resting heart rate and heart rate variability, are well studied on their own. What is missing is independent testing of the way they are combined.

What works better than looking at the score?

Asking yourself the same questions every morning and writing the answers down. A systematic review in the British Journal of Sports Medicine in 2016 examined 56 studies that measured subjective and objective markers of athlete well-being at the same time. The two kinds of measure generally did not correlate, and the subjective ones reflected both acute and chronic training load with superior sensitivity and consistency. The evidence comes from athletes rather than from forty-somethings with three children, and the subjective measures were structured questionnaires asked the same way every time, not a vague feeling. The point still stands: the cheapest instrument you own has strong support behind it.